Fine-grained Discriminative Localization via Saliency-guided Faster R-CNN

TL;DR

This paper introduces an end-to-end saliency-guided Faster R-CNN framework for fine-grained image classification, enabling automatic discriminative region localization and feature encoding, improving both accuracy and efficiency.

Contribution

It presents a novel joint framework that combines localization and classification in one network, eliminating the need for labor-intensive annotations and speeding up the process.

Findings

Achieves state-of-the-art accuracy on CUB-200-2011 dataset.

Reduces localization and classification time compared to two-stage methods.

Eliminates dependence on object and parts annotations.

Abstract

Discriminative localization is essential for fine-grained image classification task, which devotes to recognizing hundreds of subcategories in the same basic-level category. Reflecting on discriminative regions of objects, key differences among different subcategories are subtle and local. Existing methods generally adopt a two-stage learning framework: The first stage is to localize the discriminative regions of objects, and the second is to encode the discriminative features for training classifiers. However, these methods generally have two limitations: (1) Separation of the two-stage learning is time-consuming. (2) Dependence on object and parts annotations for discriminative localization learning leads to heavily labor-consuming labeling. It is highly challenging to address these two important limitations simultaneously. Existing methods only focus on one of them. Therefore, this paper proposes the discriminative localization approach via saliency-guided Faster R-CNN to address the above two limitations at the same time, and our main novelties and advantages are: (1) End-to-end network based on Faster R-CNN is designed to simultaneously localize discriminative regions and encode discriminative features, which accelerates classification speed. (2) Saliency-guided localization learning is proposed to localize the discriminative region automatically, avoiding labor-consuming labeling. Both are jointly employed to simultaneously accelerate classification speed and eliminate dependence on object and parts annotations. Comparing with the state-of-the-art methods on the widely-used CUB-200-2011 dataset, our approach achieves both the best classification accuracy and efficiency.